A long-term experiment was conducted to evaluate the effect of tillage methods on mercury and methylmercury contents in the purple paddy soil and overlying water. The experiment included five tillage methods: no-tillage and fallow in winter, ridge-no-tillage, compartments-no-tillage, paddy-upland rotation and conventional tillage. The results showed that the content of total mercury in soil had the maximum value in the 10-20 cm layer of no-tillage and fallow in winter, ridge-no-tillage and compartments-no-tillage, and the enrichment effect of no-tillage and fallow in winter was especially significant. The concentration of total mercury in soil of paddy-upland rotation and conventional tillage decreased with the increase of the soil depth, and paddy-upland rotation was specifically beneficial to the migration of mercury. The distribution of soil methylmercury was similar to that of total mercury in the soil profile. The methylation ability of soil mercury in the surface and middle of the soil profile was weaker than that at the bottom, while there was an opposite trend for other tillage methods. The concentrations of dissolved mercury ( DHg) and dissolved methylmercury ( DMeHg) in the overlaying water declined with the rise of the water depth in all treatments. The content of DHg in sediment porewater was related to the value of soil total mercury, and they had the same distribution in the soil profile. The content of DMeHg and its proportion accounted for DHg in porewater owned their largest value in the 10-20 cm layer of no-tillage and fallow in winter and ridge-no-tillage, where showed the lowest value of DMeHg in porewater for paddy-upland rotation and conventional tillage. And the percentage of DMeHg in DHg in porewater grew with the increase of soil depth of the latter two methods. Noticeably, the concentration of DMeHg and its proportion accounted for DHg in porewater were both higher than the values in overlying water for all tillage methods.